Coulmas Handbook of Sociolinguistics / Part 2 - Social dimensions of Language / part 2 - 3. Varieties and Variation

3. Varieties and Variation : The Handbook of Sociolinguistics : Blackwell Referenc... Page 11 of 13

Milroy, and Pong, 1992; Li, 1994). Attempts have also been made to project the idea of social network on to past states of language (Van der Wurff, 1990). Whereas the maintenance/change model provides a framework in which past language changes may be discussed in an illuminating way, it is dubious whether the social networks of individuals who are no longer accessible to systematic observation can be adequately reconstructed.

Social network and social class

Social network and social class represent different orders of generalization about social organization. Class accounts for the hierarchical structure of society (arising from inequalities of wealth and power), whereas network deals with the dimension of solidarity at the level of the individual and his or her everyday contacts. An attempt has been made to link the two concepts together in a sociolinguistic model by using the notion of weak network ties (Milroy and Milroy, 1992).

It is evident that close-knit solidarity ties are characteristic of lower and higher social groups, and that, in the middle sectors of society, social network density and multiplexity tend to be weak. A process model of social class, such as Thomas Højrup's theory of life-modes, suggests that different kinds of social network structure do not occur accidentally, but “fall out” naturally from different lifemodes, such as those of the self-employed, of wage-earners (both poor and relatively affluent), and of professionals. A high proportion of close-knit ties on the one hand, and of loose-knit ties on the other are consequent upon the life-modes which themselves are constitutive of distinct classes. In this way, different kinds of social network can be linked to the wider organization of society, and it is suggested that these links can be explicated by considering the properties of weak as well as strong ties (for details, see Milroy and Milroy, 1992).

8 The Sociolinguistic Variable

Critics of sociolinguistics have had much to say about the social variables discussed above, but much less about the idea of the linguistic variable. This is a relatively old concept in linguistics, most familiar in the idea of the phoneme, which typically manifests itself in the form of variants known as allophones. The sociolinguistic variable is also manifested in the form of variants. It differs from the phoneme, however, in that the focus is on social variation rather than exclusively on intra-linguistic variation. Thus the range of a sociolinguistic variable does not normally correspond to that of a phoneme, as different social values may be attached to different patterns within a given phoneme and may overlap with different phonemes. The nonidentity of the sociolinguistic variable with the phoneme is not always sufficiently emphasized by investigators.

In the foregoing it has been assumed, without comment, that sociolinguistic variables are usually phonological elements. In practice, this has often been so, but the principle underlying the method is more general than this. What is important is that variants of a variable should demonstrably be variants of the same underlying linguistic element. At higher levels of linguistic organization (particularly syntax) it is difficult to meet this condition, as it is often not clear that two syntactic variants (for example, active and passive sentence forms) have the same meaning and distribution in the language system. This difficulty is discussed by (among others) Romaine (1984), Lavandera (1978), L. Milroy (1987b). It is less often pointed out that comparable problems may also emerge at very particularistic levels of subphonemic organization. For example, in British English, the glottal stop (for /t/) occurs in different positions within words: medial, final, and in some dialects syllableinitially. However, not only does the likelihood of the glottal stop differ in these different positions; the variants that it alternates with may also differ in the different positions. Furthermore, the social meaning attached by the community to these variants may also vary according to where they occur in words or syllables. Therefore a correct quantitative statement depends on isolating environments in which we can be sure that we are dealing with variants of the same sociolinguistic phenomenon. If therefore we regard glottalization as a “variable,” we must acknowledge that it is a complex variable that contains a number of subvariables within it, and it is possible that these subvariables will display different (even contrasting) patterns.

This problem of the correct input to the variable has not been widely discussed in the literature. It seems to be most prominent in what have been called “divergent dialect” studies, where the range of variation encountered is very large (for a discussion see J. Milroy, 1992: 68–75). It also happens in

http://www.blackwellreference.com/subscriber/uid=532/tocnode?id=g9780631211938... 28.12.2007

3. Varieties and Variation : The Handbook of Sociolinguistics : Blackwell Referenc... Page 12 of 13

these studies that some salient variables do not occur frequently enough for quantification throughout the whole range (for treatment of such a variable see Milroy and Harris, 1980). It should also be noted that for reasons of time (and probably also in principle) investigators cannot quantify all the variation that exists in a speech community. Selection of representative variables depends on the skill of the investigators, and nonquantitative description is also necessary for a reasonably comprehensive account.

In the final section, we are concerned with the relevance of variation studies to our conception of what constitutes a language, a dialect, or a variety.

9 Languages and Dialects as Physical Entities

As we have noted, variation studies have led scholars to question the definition of “a language” and what kind of object a language is. Linguists have generally relied on a working assumption that there exists a structured and stable entity which we can call a language or a dialect of a language. This can be accessed or described in internal structural terms, e.g., as having a “phonology,” “grammar,” and “lexis,” without reference to society – i.e., independently of the speakers who use it in their speech communities. As noted above (see section 1), it has also been usual to treat this entity as having an invariant underlying structure. Social dialectology has called into question the discreteness of these entities that we call languages, and seeks to contribute to a clearer understanding of what we actually mean when we say that we are describing a “language.” In order to characterize a “language” or any quasidiscrete variety of a language, we need to invoke sociopolitical criteria in addition to structural linguistic criteria.

Sociolinguists (e.g., Downes, 1984; Chambers and Trudgill, 1980) commonly point out that boundaries between languages cannot be wholly determined in terms of structural difference or mutual (in)comprehensibility. Several Scandinavian languages, for example, are mutually comprehensible to a great extent and some dialects of English are not readily comprehensible to speakers of other dialects. There are many areas of the world in which variability within and between languages is very great, and some situations in which speakers may not be entirely certain as to which language they are speaking (see Grace, 1990, 1992, for comments on blurring of distinctions between certain Melanesian languages in speaker usage). Similarly, there are many situations in which two or more languages are mixed. Finally, there are rapidly changing situations, especially in the genesis and development of pidgin and creole languages, in which younger generations may use markedly different varieties of the language from those of older speakers.

From a variationist point of view, a language is a dynamic phenomenon. It is appropriate to liken languages to relatively fluid and variable physical states, and to use process models rather than product or static models in describing them.

It can be suggested that discreteness of individual languages is not inherent in the nature of language as a structural phenomenon: This apparent discreteness is socially or sociopolitically imposed. French is a “language” not merely because it has a linguistic structure that differentiates it from other languages and which is peculiarly “French,” but also because its structures are recognized, prescribed, imposed, and agreed within a particular nation-state (and certain other areas formerly influenced by this nation-state). Separateness of languages is therefore largely the result of social and political processes, and among these processes language standardization is particularly important. Our tendency to think of languages as discrete phenomena is partly conditioned by the existence of standard languages, such as standard English and standard French.

This lack of discreteness in real language states is an important matter in the study of the histories of languages. In studies of language change, there are many examples of the tendency to regard a language as a physical entity. Yet it never seems to have been possible to specify purely in terms of language structure the precise point in history at which one language “became” another language. Therefore, just as it is difficult to specify discreteness of languages in space, so it is also difficult to differentiate them in time. The point in history at which one language becomes another may have more to do with political history than with linguistic differentiation.

A final point, which arises from the above discussion, is that variation studies have in many ways blurred the Saussurean distinction between synchronic and diachronic linguistics. Mesthrie (1992) has

http://www.blackwellreference.com/subscriber/uid=532/tocnode?id=g9780631211938... 28.12.2007

3. Varieties and Variation : The Handbook of Sociolinguistics : Blackwell Referenc... Page 13 of 13

coined the useful term “panchronic” to describe such an orientation. As sociolinguists study speech communities at a single point in time, their analytic work is primarily synchronic, and their quantitative statements are synchronic statements. However, the paradigm has subtly altered the relationship between historical and other forms of linguistics, in that variation in time is grouped together with variation in space and social space as one aspect of linguistic variation. Similarly, as we have noted, process models of language in society have begun to have some impact. It is clear, however, that insofar as it is concerned with linguistic change, Labov (1994) considers the theoretical content of his work to be chiefly a contribution to historical linguistics. While this is certainly Labov's theoretical position, it is not necessarily a view shared by all sociolinguists. Equally important, as we have tried to show, is the development of an integrated account of variation encompassing not only dimensions of time and geographical space, but the various dimensions of social space, such as gender, generation, status, and network structure, discussed in earlier sections of this article.

Cite this article

MILROY, JAMES and LESLEY MILROY. "Varieties and Variation." The Handbook of Sociolinguistics. Coulmas, Florian (ed). Blackwell Publishing, 1998. Blackwell Reference Online. 28 December 2007 <http://www.blackwellreference.com/subscriber/tocnode? id=g9780631211938_chunk_g97806312119385>

Bibliographic Details

The Handbook of Sociolinguistics

Edited by: Florian Coulmas eISBN: 9780631211938

Print publication date: 1998

http://www.blackwellreference.com/subscriber/uid=532/tocnode?id=g9780631211938... 28.12.2007